james w. sewall company april 12, 2000 - merrymeeting bayjames w. sewall company april 12, 2000...

James W. Sewall Company April 12, 2000

Aquatic and Upland Habitat Assessment of Merrymeeting BayPage 1

TABLE OF CONTENTS

INTRODUCTION .....................................................................................................................2

STUDY AREA DEFINITION ..................................................................................................2

FIELD METHODS....................................................................................................................2Discussion and Observations on Field Methods....................................................................3

PHOTO-INTERPRETATION METHODS...............................................................................4Discussion and Observations on Photo-Interpretation Methods............................................6

GIS/DATABASE CREATION METHODS .............................................................................61998 (upland and aquatic)......................................................................................................61956 and 1981 Upland Classification....................................................................................71956 and 1981 Classification within the Estuary..................................................................8Completion of 1956 and 1981................................................................................................8Buildings ................................................................................................................................8

DESCRIPTION OF GIS COVERAGES ...................................................................................9Coverages Provided for each of the Three Years (yy = 56, 81, or 98) ..................................9General Coverages (used for all three years).......................................................................12

PHOTO TIMES .......................................................................................................................131956 Photography................................................................................................................131981 Photography................................................................................................................131998 Photography................................................................................................................13

DISCUSSION OF TRENDS ...................................................................................................13Land Use Changes from 1956 to 1998 Over the Entire Study Area....................................14Silt and Sand Changes from 1956 to 1998...........................................................................15Vegetation Changes from 1956 to 1998 Over the Entire Bay.............................................19Vegetation Changes from 1956 to 1998 by Sub-Section.....................................................20

SUMMARY.............................................................................................................................24

REFERENCES ........................................................................................................................26



INTRODUCTION

In early 1998, the Friends of Merrymeeting Bay (FOMB) commissioned James W. SewallCompany of Old Town, Maine and Kleinschmidt Associates of Pittsfield, Maine to providetechnical support for aerial photography, mapping, and analysis of trends in the Bay andsurrounding half-mile upland buffer area. The study would be the latest of several suchstudies of the Bay proper, but the first to include the neighboring uplands. Previous studieswere conducted in 1956, 1961, 1966, and 1981 (Spencer, 1966; Anderson, 1982). Thepurpose of the present study was to document conditions in the study area in 1998 and reporton trends in vegetation and land use since 1956. In order to do this, FOMB wished toreplicate the earlier vegetation classifications inside the Bay and, in addition, map uplandland use and land cover. Classifications were carried out by photo interpretation of 1956,1981, and 1998 aerial photography. Mapping and analysis were done using GeographicInformation System (GIS) techniques. The following discussion of the study and its resultsis divided into four sections, three of which present and discuss different aspects of themethodology (field method, photo interpretation, and GIS). The final section presentssummary figures, tables, and some discussion on change. Maps, ARC/INFO GIS coverages,an ARCVIEW summary project, and a comprehensive data table in Excel spreadsheet formataccompany this report on CD-ROM and may be used as an extension to it.

STUDY AREA DEFINITION

The 1998 photos were classified for upland types in an area within 1/2 mile of the Bay and itsmajor tributaries to head of tide. For the purpose of defining the 1/2-mile buffer, the 1:24000USGS map of the Bay and its tributary rivers (Androscoggin, Muddy, Cathance,Abagadasset, Eastern, and Kennebec) was used. The 1998 photos provided coverage of theBay and to the head of tide on all major tributaries except the Kennebec, which was coveredonly to the north end of Swan Island. A small portion of the 1/2-mile buffer along thewestern side of the Eastern River was also not covered by the 1998 photography. A GIS file(mmb98covarea) shows the effective study area (the intersection of the study area as definedby the above mentioned 1/2-mile buffer and stereo photo coverage). For the trend analysis,the study area consisted of the area common to the photo coverages of 1998, 1981, and 1956(Figure 1).

FIELD METHODS

Field work involved ground truthing of preliminary photo interpretation work using 1998color infrared (CIR) aerial photos, and data gathering for detailed reference wetland work.This fieldwork was completed during the last few days of July and the first half of August1999 to coincide with the phenology of the aerial photographs. The study area was alsodivided into subsections corresponding as much as possible to those defined in earlier studies(Anderson, 1982). A portion of each sub-section was visited. Fieldwork was also used to



establish or confirm aerial photo signatures for the various marsh types and substrates (silt,sand, rock). Out of a total of 13 person-days of fieldwork, only a portion of one day wasdedicated to ground truthing of upland habitat types. The vast majority of the time was usedto establish reference wetland data and to ground truth marsh types. A summary of theresults of the reference wetland study, including species detail and photographs ofrepresentative stands, may be found in the separate “Wetlands Document,” prepared inSeptember 1999 by Alan Haberstock of Kleinschmidt Associates.

Discussion and Observations on Field Methods

• Timing of fieldwork and aerial photography dates is important. Many marshes appear tobe dominated by species other than wild rice through June and even into July. Wild riceis an annual and, in the beginning part of the growing season in mixed marshes, it isrelatively inconspicuous beneath the foliage of perennial vegetation. Perennial speciesexhibit quicker starts due to stored energy but are overtopped by wild rice by July.Therefore, many wetlands were called “mixed” (no single species had an areal coverageof 70% or more) that might have been called a single-species wetland other than wild rice(such as pickerelweed or soft-stemmed bulrush) in the beginning part of the growingseason. Many of the “mixed” marshes contained a dominance of wild rice (usuallyaround 25-70% of the total area coverage attributable to wild rice) but not enough toreach the 70% threshold.

• Exotic/invasive species such as purple loosestrife (Lythrum salicaria) and common reed(Phragmites australis) are not prevalent in the Bay. Purple loosestrife occurs in smallpatches throughout the Bay near the high water mark. This species could affect theecological integrity of the Bay if it increases. It is displacing native, water-edge specieslike cardinal flower (Lobelia cardinalis) and blue vervain (Verena hastata), not emergentspecies like pickerelweed and wild rice. Common reed is not a problem at present in theBay.

• We have observed that Merrymeeting Bay is dynamic. There are wetlands that havechanged shape and extent between 1981 and 1999, particularly at the mouth of rivers andin the central part of the Bay. Some of the marsh complexes out in the middle of the Bayseem to have even changed shape a bit between ‘98 and ‘99. The sediment dynamicsresult in shifting wetland locations and coverage. Wild rice in fact may thrive in part dueto these shifting sediments, easily seeding-in where the sediment is just right and whereperennial species may get smothered (in depositional environments) or removed(erosion). Factors such as ice dynamics in the winter may also favor pioneer species dueto physical action of the ice on substrates.

• Wild rice CIR signatures (i.e., tone, texture, color) varied from stand to stand dependingon how robust the stand was whereas other species had more uniform signatures.

• Perhaps the most dynamic and ubiquitous species in the Bay is wild rice. This annual,non-persistent, wind-pollinated, emergent grass grows to a height of more than 2 meters.It favors soft, muddy or silt areas where there is at least some movement of water and



little competition from other species (at least some patches of bare substrate). Thisspecies does not like stagnant water. Its survival into the next year depends entirely on itsyearly seed crop. The population of wild rice in the Bay can swing from year to year inresponse to many variables including disease (certain fungi occasionally infect the plant),weather, sediment dynamics, seed consumption, and even wind direction duringpollination, water flows during germination and seed dispersal, and carp numbers andhabits (carp like to uproot and eat young plants). Spencer (1966) points out that thevariability of wild rice can cause problems for photo-interpretation and for trend analysis,because “in years of good crops a dense stand of rice sometimes forms an overstory thatprovides the photo image but which may obscure a moderately dense understory of suchplants as yellow water lily.”

PHOTO-INTERPRETATION METHODS

Three dates of photos were interpreted: 1998, 1981 and 1956 (all in August). The 1998photos were color infrared (CIR) film positives at a scale of 1"=1000'. The 1981 photos weretrue color prints at a scale of 1"=660'. The 1956 photos were true color positivetransparencies, also at a scale of 1"=660'. All photo-interpretation labels were a singlealphabetic code that would allow expansion by computer routine into component codes ofland use class, land use sub-class, vegetation, and (in 1998) forest cover type.

Upland land use was classified using a modified version of Anderson’s (1976) classification,with alphabetic rather than numeric codes and with sub-classes added as deemed necessary.In 1998 only, forested areas were broken down into softwood, mixed wood, and hardwood aswell as typed for height and density. In other years, the forest land use type was notsubdivided into subclasses. Wetlands were classified as “WET” under land use class andalso typed under the Cowardin et al (1979) wetlands classification system. The Bay’semergent vegetation areas were sub-classified for the same target species classes as wereused in previous studies (cf. Spencer, 1966). A summary of all codes is given in a latersection of this document (DESCRIPTION OF GIS COVERAGES).

In classifying emergent vegetation, a vegetated area was given a single species code if thatspecies covered 70% or more of the area. Density of the coverage had to be at least 30% foran area to be considered vegetated. Areas with lower vegetation density than 30% wereclassified in terms of the substrate (silt, sand, rock, open water). If there was less than 70%coverage of one species it was classified mixed. If a wetland had a species other than thetarget species that had coverage of 70% or more of the vegetated area, the wetland was alsoclassed as mixed. For example, three square bulrush was a species that sometimes formedsmall monocultures (stands with 70% or more of the total area). These areas were put in themixed emergent class.



The upland classification was modified on the 1981 and 1956 true color photos. The forestwas not broken down into general composition class or height and density for these years.Otherwise, uplands were classified to land use as in 1998. The wetland areas were alsoclassified the same as on the 1998 photos. A basis for truthing of the aquatic vegetationportion of the classification was provided by annotated 1957 black-and-white photography inDavid P. Olson’s MS thesis (Olson, 1958).

The 1956 and 1981 photos did not cover the entire project area. Figure 1a shows therelationship between the three years’ effective study areas, and the area where they overlap.This is the study area referred to in all summary tables in the trend analysis section of thisreport. Figure 1b shows the subsections used in this study. Although the subsections in earlierstudies did not extend into upland areas, they have been extended in this study for purposesof reporting upland land use trends by subsection. The lines from the earlier studies wereextended so that upland areas were also assigned to a sub-section. This allows analysis notonly of changes in the bay, but also of the land use changes by sub-section.

Figure 1. a) The study area a defined by the intersection of three years of photocoverage, and (b) the definition of subsections.

a.

b.



Discussion and Observations on Photo-Interpretation Methods

• No definite limit of the Bay was defined by the previous studies. The limit used for 1998corresponds to the area coded as “YES” in the Estuary attribute of each of theMMByy_ALL coverages and is approximately 10,700 acres in all three years. This limitincludes the inter-tidal and sub-tidal areas as well as these could be determined, and alsoincludes the emergent vegetation areas of target species adjacent to the Bay.

• The photography varied among different years. 1998 photos were at a smaller scale andcolor infrared; 1956 photos were badly faded, true color transparencies, and 1981 weretrue color prints that varied in quality over the entire area but were somewhat better inquality than 1956.

• Aquatic beds are submerged even at low tide (see also Spencer, 1966). The highresponsiveness of CIR film to vegetation probably makes aquatic beds more visible(1998) than on true color film (1956, 1981).

• Of the three years of photography used in the present study, 1998 had the lowest tide atthe time of photography, followed by 1956, then 1981 (see further data under TRENDANALYSIS).

• The methodology of previous photo-interpretation of the Bay was to interpret spots undera dot grid (H. E. Spencer personal communication to L. B. Feero). This method hasdifferent properties than the polygon method used in the present study. What was beinginterpreted in the earlier studies was the condition that fell immediately to the upper leftof each dot. Every such spot was interpreted, dots were counted by category, and acreagedetermined for the category by multiplying the category count by the dot acreage factor.For example, on the 1"=660' photography of the earlier studies, each square inch covers10 acres. In a 16 dots/ inch grid, each dot would have represented .625 acres. Eventhough the minimum area of the present study’s polygons is 1/8 acre, the polygon methodmay possibly increase the “mixed emergents” category for complicated mosaics ofspecies that could have been separated into individual species in the dot method.

GIS/DATABASE CREATION METHODS

1998 mapping was done first and used as a base for 1956 and 1981.

1998 (upland and aquatic)

• Map features used for control in transferring photo-interpreted data from the 1" = 1000'photo base were plotted manuscripts at the same scale from USGS 7.5' maps on CD-ROM (georeferenced raster images copyrighted by Land Info International). For the Bayregion, these maps are of mid- to late-seventies origin.



• Photo-interpreted type boundaries were transferred from photos to the control base usinga vertical sketchmaster. This device is suitable for adjustment of minor scale differencebetween source photo and control base in relatively flat terrain. When transferring fromphotos to controlled base, significant shoreline features, buildings, and road intersectionswere used for control. This method does not eliminate distortion due to reliefdisplacement and scale changes within photo frames, but does allow these variations tobe controlled by the matching of photo features to their counterparts on the control base.The USGS base provided good control except within the Bay proper, where control detailwas sparse. In these areas, adjustments were made between photos to make them fit eachother and the boundaries of the bay.

• The transferred lines were checked and then scanned on a TRUSCAN model CS400/10scanner to .TIF format.

• ARC/INFO GRID software was used to convert .TIF to geo-referenced ARC/INFO lineformat.

• Lines were edited and label points added to create error-free polygon topology.

• Polygons were assigned attributes using the photo-interpreter’s code.

• A 100% check was made to ensure completeness of labeling and an accurate match to thephoto-interpreter’s classification.

• The polygons were digitally overlaid with US Fish and Wildlife Service NationalWetlands Inventory (NWI) maps to supply major wetlands system classes for eachpolygon. This established the breaks between Riverine and Estuarine systems inessentially the same locations as in the NWI.

• A computer reclassification routine was executed to populate all attributes base on thephoto interpretation code.

1956 and 1981 Upland Classification

• Lines and labels of major 1998 upland land use types as well as roads and water wereplotted at 1"=660' scale (same as photos) on clear inking film.

• A photo-interpreter overlaid these maps on the 1956 or 1981 photos and interpretedupland land use directly to the clear film overlay. The overlay was shifted as necessaryon top of the photo to match major features and adjust for scale variation across thephoto. Land use boundaries that were changed (from the 1998 base) were inked and linesneeding removal were cross-hatched for later deletion. The rule of thumb used todetermine whether to change the line was a minimum variation in a boundary of onetenth of an inch. This is equivalent to 66 feet at ground scale. Any lines on the 1998base that overlaid a corresponding land use boundary on the photo within a tenth of aninch were left in place. This method was used to avoid creation of spurious sliverpolygons and ensure that change was truly change and not variation due to the horizontalpositional inaccuracy of the transfer method.



• Using these edited maps as a source, the 1998 polygon file was edited at a digitizingworkstation to create 1956 or 1981 upland land use, deleting or adding lines andmodifying polygon labels as necessary to create a new polygon layer appropriate to theyear.

1956 and 1981 Classification within the Estuary

Because of the great amount of difference in the Bay’s water and vegetation polygonconfigurations between dates, the editing procedure used for uplands was not used for thearea within the Bay. Instead, complete new sets of lines were created for aquatic typepolygons (vegetation and other wetlands types including open water). The same method aswas used for the original 1998 mapping was used to capture polygons, except that controlwas provided by the already-created 1998 land use types instead of the USGS raster base(same base maps used in the land use transfer). The steps were:

a) Transfer from photo to map using vertical sketchmaster with detail on 1998 landuse base providing control.

b) Scan mylar overlays to .TIF format.c) Automate and convert to geo-referenced ARC/INFO line format.d) Clean up lines and add label points to create polygon topology.e) Assign photo interpreter's classification code to polygons.f) Quality control to ensure completeness of labeling and accurate match to photo-

interpreter’s classification.

Completion of 1956 and 1981

• Aquatic types were merged with upland land use. The aquatic layer superceded theupland layer in any area of overlap.

• Adjustments were made along the border between the upland land use and aquatic typesto eliminate slivers. Routines were used to identify polygons that bordered sliverpolygons and assign the attribute of the major adjacent type. These assignments werechecked before dissolving borders between polygons with like labels.

• The coverage was intersected with NWI mapping and dominant wetlands systemcomputed for each. A computer reclassification routine was executed to populate allattributes based on the photo interpreter's classification.

Buildings

Buildings were first screen-digitized to a shape file in ARCVIEW from the USGS raster base(copyright Land Info International). This provided a basis for building locations in all butheavily developed areas (municipalities, trailer parks) where individual buildings are notshown by USGS. This USGS building location coverage was plotted on the clear film base



that was also used for control in 1956 and 1981 land use mapping (see above). These clearmylars were overlaid on the 1956 and 1981 photographs and buildings not present (i.e. notvisible on photographs) were marked for deletion, or new buildings visible on thephotographs were marked for addition. A similar process was used for 1998, with mapsplotted at the 1" = 1000' scale. Building points that originated from photographs were copiedforward between years. The result of these procedures is that a building present in multipleyears has identical coordinates in all years. All buildings except small outbuildings weredigitized. No attributes were applied to the ARC/INFO point features representing thebuildings.

DESCRIPTION OF GIS COVERAGES

• All GIS coverages are in ARC/INFO 7.2.1 format. In addition to being provided asARC/INFO coverages accessible from ARCVIEW they are present on the delivery CD asARC/INFO export (E00) files. The CD also contains metadata for land use/land cover,building, and base feature coverages.

• The projection used for all coverages is UTM Zone 19, NAD27, meters.

Coverages Provided for each of the Three Years (yy = 56, 81, or 98)

Coverages MMByy_ALL - this is the primary coverage type of the study, containing all theattributes for land use, wetlands, and vegetation. Attributes are the same for the three yearsexcept that 1998 has enhanced information on the forest type, adding four attributes. Thefollowing descriptions of attributes does not include ARC system attributes, but these arepresent. The non-system attributes and code values are as follows:

ALPHA - Unique photo-interpreted code from which other codes (except WETL_SYS)were derived. For example, the letter "A" (in early phase "PEM/A") was used by thephoto-interpreter to mean wild rice. This value was recoded to WET for land use, EM forland use subclass, and EM1 for WETL_CLS. Similar recodings were done for othervalues of ALPHA. ALPHA itself was never changed from the original interpreted valueunless changed by the interpreter. Thus, it could always be checked against the acetatephoto overlays on which interpretation codes were written.

LUCLASS Land use class 3-character alphabeticAGR Agricultural (cropland or pasture in current or recent use)COM Commercial (business or commercial predominant use)FOR Upland Forest (predominant land use forest)IND Industrial (manufacturing facilities and associated land)OW Open Water (lake, river, or bay areas under water greater than a few inches

deep at the time of photography)RES Residential (homes and related neighborhoods)USS Upland Scrub-Shrub (usually abandoned agricultural field)UHE Upland Herbaceous (large expanses of lawn not obviously associated with

residence)



WET Wetland (general class for upland and bay wetlands other than open water)LUSUBCLASS - Land use subclass. 4-character alphabetic. Some of the land use classeswere broken into subclasses. Those that were not broken down were simply repeated inthe subclass attribute. This allows the land use subclass attribute to be used as anexhaustive alternative breakdown of land use. The sub-codes, by major code are asfollows:

AGR, RES, IND no subclass breakdown, repeated primary code.COM either COM or one of the following alternate codes

YD harvest yarding area in forestGP gravel pitTL* primary use transmission line

*Transmission lines were LUCLASS ‘COM’ and LUSUBCLASS ‘TL’ code only when they crossedforested land. If land under a transmission line was being used for agriculture, it was typed AGR.

WET no subclass breakdown for uplands in 1956, 1981AB aquatic bedEM emergent vegetationPFO forested wetlandHT high tide zone – unvegetated areas above high tideROCK rock showing within the bay or bay areaSAND sand flats (may be submerged)SILT silt (may be submerged)PSS palustrine scrub-shrub wetland

FOR 1998 only broken into sub-classes - other two years use FOR1998 sub-classesSW softwoodMI mixed softwood and hardwoodHW hardwoodLG ledge or rock outcrop in upland area

OW, USS, UHE no subclass breakdown, repeated primary code

WETL_SYS - Wetlands system. One-character alphabetic code for ecological system,under the scheme of Cowardin et al (1979). These codes were assigned by overlayingwith US Fish and Wildlife National Wetlands Inventory (NWI) maps, then assigning thedominant code of the US Fish and Wildlife Service’s National Wetlands Inventory maps.Thus, for example, the breaks between Riverine and Estuarine in the bay roughly matchthe breaks used on the NWI maps.

E Estuarine (not subdivided into tidal and subtidal)L LacustrineP PalustrineR RiverineU Upland



WETL_CLS. An approximation of the ecological subsystem from the Cowardin scheme,using the following codes. 3-character alphabetic.

AB submerged vegetationEM1 persistent emergent vegetationFO forestedHE herbaceousRS rocky shoreSB4 streambed sandSB5 streambed siltSS scrub-shrubUB unconsolidated bottom

VEG and VEG_DESC - Vegetation code. 1-character alphabetic code for emergentvegetation type and 20-character description with code. Codes are as follows:

A wild riceB yellow water lilyC sweet flagD softstem bullrushE river bullrushF pickerelweedG mixed emergentsH broad-leaved cattail

FORCOV - 1998 only. Forest cover types are made up of three components, speciescomposition, size class, and crown closure. The four-character FORCOV code is madeup of the following codes for each of these components:

FORSPP (species composition)

H Greater than 75% HardwoodS Greater than 75% SoftwoodHS Hardwood-dominated mixture of hardwoods and softwoods (hardwood 50-75%)SH Softwood-dominated mixture of hardwoods and softwoods (softwood 50-75%)

FORSIZ (a size classification roughly indicative of commercial product or seral stage)

1 Seedling (



ESTUARY – “YES” if polygon is part of the area considered to be estuary, otherwiseblank. This was determined by taking all emergent vegetation types bordering on theBay.

UPL – “Yes” if polygon is part of the area considered to be upland, otherwise “NO”.

Coverage MMBXXALL - master intersection coverage containing linking -ID numbers ofthe three years' primary coverages. This coverage was used in ARC/INFO to provide thecross tabulation of all years that is presented in the Excel table DATA.XLS.

Coverages MMByyROADS - modified USGS roads coverage (Source: Maine Office ofGIS) within the coverage area for each year to reflect road status at the time of thephotography. The only non-system attribute of this coverage is:

CLASS - Numeric, single-byte integer

1 Interstate2 Primary3 Secondary4 Improved5 Unimproved6 Trail

Coverages MMByyCOVAREA - stereo photo coverage area within study area. Anadditional coverage, MMBXXCOVAREA, is provided which describes the stereo coveragearea common to all three years’ coverage.

Coverages MMByyBLD - a point coverage for each of the three years representingbuildings. No attributes were used other than system point attributes. MMBXXBLD is arelated coverage intersecting the building coverages with the MMBSECTIONS coverage sothat numbers of buildings could be tabulated by section.

General Coverages (used for all three years)

Coverage MMBTWPS - Township boundaries for towns overlapping the study area(source: Maine Office of GIS). The attributes of this coverage are:

TOWN - 20 character alpha, name of Maine civil division (city, town or township).GEOCODE - 5-character Maine code for civil division.

Coverage 24KROADS - Same as MMByyROADS except these are the same as the USGSroads, which were mapped in the mid- to late-seventies. This is used to provide context andconnectivity for the roads within the study area (source: Maine Office of GIS). Attributes:

CLASS - same codes as for MMByyROADS coverages (see above).



Coverage MMB_COAST - USGS polygonal water features covering the study area (source:Maine Office of GIS). The attributes are:

RIVER - single digit integer, 0 if island, 1 if not.

Coverage MMB_RIVERS - USGS polygonal water features covering the study area(source: Maine Office of GIS). The attributes are:

ISLAND - single digit integer, 1 if island, 0 if not.

Coverage MMB_STREAMS - USGS linear water features covering the study area (source:Maine Office of GIS).

Coverage MMB_MASK - One-mile buffer on study area used to mask external water androad detail in ARCVIEW map layouts. The attribute here is INSIDE, which has a value of100 for areas inside the one-mile buffer and 1 for areas outside.

PHOTO TIMES

The times (eastern standard time) of photography for the three dates were as follows:

1956 Photography

Flight Lines 1-8 taken on August 25 between 10:00 am and noon.Flight Lines 9-13 taken on August 27 between 11:00 am and 1:00 pm.

1981 Photography

Taken on August 18 at a mean time of 9:00 am.

1998 Photography

Taken on August 9 at a mean time of 8:50 am. The actual flight line times were:

Line 1 9:12 to 9:13 Line 7 8:26 to 8:31Line 2 9:08 to 9:09 Line 6 8:34 to 8:39Line 3 9:01 to 9:05 Line 5 8:43 to 8:47Line 4 8:52 to 8:57

DISCUSSION OF TRENDS

While a detailed analysis of the data is beyond the scope of this report, some comment on thenature of the study’s products and their appropriate uses is appropriate. The discussionbelow first deals with overall trends in land. Then follows a brief analysis of tide levels atthe different study dates and how their influence can be factored out of the analysis of trendsin the intertidal zone. Finally, some of the more significant trends in the Bay and itssubsections are identified.



Land Use Changes from 1956 to 1998 Over the Entire Study Area

Increases were seen in forested land (2198 acres), residential land (1789 acres), andcommercial land (331 acres). Decreases were observed in agricultural land (2413 acres) andabandoned field land or upland scrub-shrub (2497 acres). The term “abandoned field” isfrequently used in this paper and on maps as a synonym for the “upland scrub-shrub” type(code USS). Table 1 summarizes the changes that have occurred from 1956 to 1998.

Table 1. Merrymeeting Bay Land Use Summary 1956-19981956 1981 1998

Agriculture (AGR) 4,672 2,611 2,259Commercial (COM) 498 913 829Forested (FOR) 12,020 14,539 14,219Industrial (IND) 7 27 64Residential (RES) 747 1,547 2,537Abandoned Field (USS) 2942 929 445Wetland (WET) 6,363 5,954 7,315Open Water (OW) 5,695 6,412 5,245Total 32,945 32,945 32,945

A second set of tables (2a. and 2b) show the above categories for 1956 cross-tabulatedagainst the same classes for later years show the types of transitions that occurred. Thetables use the codes for the classes listed above. Numbers in columns represent 1956categories and how they are reclassified in the 1981 and 1998 maps. Similarly, numbers inrows represent the 1956 origins of 1998 and 1981 classes. For example the 1998 residentialclass (total 2537 acres) had its origin in 1956 classes AGR (892 acres), COM (29 acres),FOR (657 acres), RES (645 acres), USS (298 acres), WET (11 acres), and OW (4 acres).(The smaller categories (COM, WET, and OW) may be discounted as probably due to theerror inherent in the mapping process).

Table 2a. Land Use Change from 1956 to 1998Land Use 56

Land Use 98 AGR COM FOR IND RES USS WET OW Grand TotalAGR 2,029 3 72 19 114 22 0 2,259COM 97 373 239 15 99 4 3 829FOR 1,221 79 10,437 58 2,167 201 56 14,220IND 13 3 25 2 9 12 64RES 892 29 657 645 298 11 4 2,537USS 320 0 18 2 96 9 0 445UHE 18 3 1 5 5 32WET 80 7 500 6 143 5,458 1,122 7,316OW 3 1 71 3 12 646 4,509 5,245Grand Total 4,672 498 12,020 7 748 2,942 6,363 5,695 32,945



Reading the AGR column of table 2a. similarly shows 1998 types for land typed asagriculture in 1956. About 43% remained in agriculture (2029 acres), but 26% (1221 acres)became forest, 19% (892 acres) became residential, about 7% (320 acres) became abandonedfield (USS), and about 2% (97 acres) was put to some commercial use. A comparison oftables 2a and 2b shows that most of the transition of agricultural land to other uses hadoccurred by 1981, when only about 50% of the 1956 agricultural land remained inagriculture. Other trends can be similarly analyzed using these tables.

Table 2b. Land Use Change from 1956 to 1981Land Use 56

Land Use 81 AGR COM FOR IND RES USS WET OW Grand TotalAGR 2,329 12 101 21 115 33 0 2,611COM 139 371 237 16 131 19 1 913FOR 1,036 91 10,738 0 74 2,254 293 61 14,548IND 8 7 12 27RES 449 17 326 596 152 5 3 1,547USS 644 5 94 30 147 8 1 929UHE 11 0 1 1 13WET 60 2 440 4 120 4,954 412 5,993OW 4 1 76 6 10 1,050 5,217 6,364Grand Total 4,672 498 12,020 7 748 2,942 6,363 5,695 32,945

Silt and Sand Changes from 1956 to 1998Although all aerial photographs were taken at or near low tide, there was still some variationin tide level between the three years. Before any conclusions could be drawn regardingchanges in sand and silt, these differences needed to be evaluated. A close inspection wasmade of the photography for each year in order to determine the relative tide differences.Several spots were chosen within the bay that had highly visible rocks close to the surface ofthe water. At each spot, a particular rock was inspected for each year to determine relativetide heights. Figures 2 (below Chops, east side of Lines Island) and 3 (off south-west tip ofBrick Island in Middle Bay) show two spots that were examined. 1998 photography provedto be the lowest tide level, followed by 1956, then 1981, which is reflected in the subtidalopen water acreage (5194, 5683, and 6339 acres of open water respectively). Total acreageof the estuary is roughly the same for each year (about 10,700 acres).

Table 3. Intertidal and Subtidal Area by Year1956 1981 1998

Acres Percent Acres Percent Acres Percent

Total Intertidal Zone 4,990 47% 4,324 41% 5,530 52%Total Subtidal* Zone 5,683 53% 6,339 59% 5,194 48%Total Area 10,673 100% 10,663 100% 10,724 100%*Eliminating open water outside the estuary makes this number lower than the "land use" open water number.



Figure 2 Chops

Figure 2. Samerock in Chops sub-section off the eastside of Lines Islandused to determinerelative tide levelsamong three timeperiods. 1956

Medium Tide Level

1981Higher Tide Level

1998Lower Tide Level



Figure 3 Middle Bay

Figure 3. Samerock in Middle Baysub-section nearsouthwest tip ofBrick Island used todetermine relativetide levels amongthree time periods.

1956Medium Tide Level

1981Higher Tide Level

1998Lower Tide Level



In silt and sand areas, the water line is very difficult to discern, but there is usually a clearbreak at the edge of the silt or sand where deeper water appears to begin. Silt and sand weretherefore interpreted when they could be distinguished even if submerged. A review of theclassification was conducted after all areas were interpreted to check for consistency ininterpretation of sand and silt. As a result of the review, the photo interpreter felt the depthof water and therefore the amount of submerged silt and sand were relatively consistent fromyear to year. That is, about the same amount of underwater silt and sand is interpreted ineach case.

Although based on total acreage, the entire bay saw an increase (1338 acres) of silt and sandfrom 1956 to 1998, i.c. More sand and silt was identified in the 1998 interpretation due tothe lower water level. To correct for the difference in acres of sand and silt due to waterlevel, the decrease in open water within the estuary from 1956 to 1998 was calculated. Therewas a decrease of 489 acres in open water. Correcting for this difference, silt and sandincreased by about 850 acres between 1956 and 1998. (The same trend was observed byK.H. Anderson in 1982 who noted a 646 acre increase in silt and sand between 1956 and1981.) Figure 4 shows that total silt, sand, and water were roughly equal between 1981 and1998, but greater than in 1956.

Figure 4. Breakdown of Vegetation and Water Components

Breakdown of Vegetation and Water Components

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

1956

1981

1998

Wild Rice

Mixed Emergents

Subm. Aquatics

Sweet Flag

Softstem Bulrush

River Bulrush

Pickerelweed

Yellow Waterlily

BdLeaved Cattail*

Sand

Silt

Open water



Vegetation Changes from 1956 to 1998 Over the Entire Bay

Vegetation changes are reported in acres as well as a percent of change within each class,calculated as (1998acres-1956acres)/1956acres*100. Major increases for individualvegetation classes were seen in wild rice (283 acres or 30% increase) and submerged aquatics(102 acres or 30%). There was a major decrease in mixed emergent vegetation (-1117 acresor 49% decrease) possibly attributed to the dominance of wild rice at the time of photographyin 1998, which decreased the mixed class. There was also a decrease in yellow waterlily (-101 acres or 84% decrease) and pickerelweed (-65 acres or 37% decrease).

Overall, the data show a change toward greater quantities of wild rice in 1998. It appearsthere may be a reciprocal relationship between wild rice and mixed emergents. When wildrice is an important component of a marsh stand and growing vigorously, it may easilyachieve the 70% of cover needed to call the stand wild rice. When it does not have as good aseason, the same area might well fall into the mixed emergent or other class. But in 1998,the wild rice dominance was clear where it was the interpreted species.

Although there was an increase in wild rice, there was a decrease (-786-acre or 23.6%) inemergent vegetation throughout the estuary. This difference had already occurred by 1981,with no reduction in emergent vegetation between 1981 and 1998. (The 1956 to 1981reduction is corroborated by Anderson (1982), who noted a 650-acre decline in totalvegetation in the Bay between 1956 and 1981). See Table 4 for a summary of estuarinecover composition change. For composition breakdown, see Figure 4.

Table 4. Intertidal Zone Composition ChangeComposition Change Percent change

1956 Acres 1981 Acres 1998 Acres 1956 to 1981 1981 to 1998 1956 to 1998

Wild Rice 933 959 1,216 3% 27% 30%

Mixed Emergents 2,278 1,770 1,161 -22% -34% -49%

Subm. Aquatics 336 178 438 -47% 146% 30%

Sweet Flag 8 4 3 * * *

Softstem Bulrush 221 206 233 -7% 13% 5%

River Bulrush 13 9 28 * * *

Pickerelweed 178 124 113 -30% -9% -37%

Yellow Waterlily 121 50 20 -59% -60% -84%

BdLeaved Cattail 4 7 95 * * *

Total Vegetation 4,092 3,306 3,306 -19% 0% -19%

High Tide Zone 52 7 44 * * *

Rock 8 30 4 * * *

Sand 221 346 695 57% 101% 215%

Silt 617 635 1,481 3% 133% 140%

* No change (absolute size of the category too small to conclude relative change)



Vegetation Changes from 1956 to 1998 by Sub-Section

Swan Island

There were major increases in wild rice (233 acres), and silt (149 acres) in the Swan Islandsub-section. Mixed emergent vegetation decreased (-414 acres), as well as submergedaquatics (-56 acres), softstem bulrush (-17 acres) and yellow waterlily (-12 acres). There wasan overall decrease (-194 acres) of emergent vegetation.

Eastern River

This subsection saw an increase in wild rice (83 acres) and silt (55 acres). The mixedemergent vegetation decreased (-117 acres), as well as softstem bulrush (-41 acres) andpickerelweed (-19 acres). There was an overall decrease of emergent vegetation (-90.5acres).

Eastern River Vegetation Change

-117

-41

-19

-1

4

55

83

-150 -100 -50 0 50 100

Mixed Emergents

Softstem Bulrush

Pickerelweed

Subm. Aquatics

BdLeaved Cattail

Silt

Wild Rice

Acres Changed

Swan Island Vegetation Change

-414

-56

-17

-12

-2

9

9

11

149

233

-500 -400 -300 -200 -100 0 100 200 300

Mixed Emergents

Subm. Aquatics

Softstem Bulrush

Yellow Waterlily

River Bulrush

BdLeaved Cattail

Pickerelweed

Sand

Silt

Wild Rice

Acres Changed



Middle Bay

The only large-scale increase in this subsection was silt (197 acres). Major decreases wereobserved for mixed emergent vegetation (-143 acres), wild rice (-50 acres), and submergedaquatic vegetation (-41 acres). There was an overall decrease of emergent vegetation (-175acres).

Abagadasset River

This sub-section saw the largest increase in silt (302 acres) and sands (203 acres). Therewas an increase in wild rice (34 acres) and broad-leaved cattail (18 acres). Decreases wereobserved in mixed emergent vegetation (-103 acres), yellow waterlily (-13 acres), submergedaquatic vegetation (-8 acres). There was a decrease in emergent vegetation (-72 acre).

Middle Bay Vegetation Change

-143

-50

-41

-2

3

20

197

-200 -150 -100 -50 0 50 100 150 200 250

Mixed Emergents

Wild Rice

Subm. Aquatics

Yellow Waterlily

Sand

River Bulrush

Silt

Acres Changed

Abagadasset River Vegetation Change

-103

-13

-8

-8

-5

1

2

18

34

203

302

-150 -100 -50 0 50 100 150 200 250 300 350

Mixed Emergents

Yellow Waterlily

Subm. Aquatics

Pickerelweed

Sweet Flag

River Bulrush

Softstem Bulrush

BdLeaved Cattail

Wild Rice

Sand

Silt

Acres Changed



Muddy and Cathance Streams

Silt increased (129 acres) in this sub-section. Other increases were observed in wild rice (46acres), broad-leaved cattail (29 acres), softstem bulrush (29 acres), and submerged aquaticvegetation (15 acres). Mixed emergent vegetation decreased (-81 acres), as well as yellowwaterlily (-72 acres), and pickerelweed (-32 acres). There was an overall decrease ofemergent vegetation (-83 acres).

Androscoggin

This sub-section saw the largest increase in sand (253 acres), and submerged aquaticvegetation (196 acres). Increases were also observed in wild rice (92 acres) and softstembulrush by 65 acres. Decreases were observed in mixed emergent vegetation (-347 acres),silt (-35 acres) and pickerelweed (-23 acres). There was a decrease in emergent vegetation(-202 acres).

Androscoggin Vegetation Change

-347

-35

-23

4

5

65

92

196

253

-400 -300 -200 -100 0 100 200 300

Mixed Emergents

Silt

Pickerelweed

Yellow Waterlily

BdLeaved Cattail

Softstem Bulrush

Wild Rice

Subm. Aquatics

Sand

Acres Changed

Muddy/Cathance Streams Vegetation Change

-81

-72

-32

-1

-1

3

15

29

29

46

129

-100 -50 0 50 100 150

Mixed Emergents

Yellow Waterlily

Pickerelweed

River Bulrush

Sweet Flag

Sand

Subm. Aquatics

Softstem Bulrush

BdLeaved Cattail

Wild Rice

Silt

Acres Changes



Chops

Increases in mixed emergent vegetation (88 acres), silt (67 acres), and broad-leaved cattail(25 acres). There were decreases in wild rice (-156 acres), softstem bulrush (-25 acres), andyellow waterlily (-7 acres). There was a decrease in emergent vegetation (-71 acres).

Buildings

The digitized buildings were tallied for the study area, and by sub-section. The numbers ofbuildings increased by 235 over the whole study area, from 1956 to 1981, and increasedanother 1283 from 1981-1998. When divided by sub-section, several areas decreased innumber of buildings from 1956 to 1981. This may be attributed to the decline of agriculturalbuildings, which were then replaced by residential building after 1981. See Table 5 andFigure 5 for a summary of sub-section changes.

Table 5. Numbers of Buildings in Study Area by Section and YearSection 1956 1981 1998

Swan Island 577 563 693Eastern River 104 100 162Abagadasset R. 55 50 90Middle Bay 44 35 41Muddy/Cathance 105 110 189Chops 254 245 375Androscoggin R. 715 986 1822Total 1854 2089 3372

Chops Vegetation Change

-156

-25

-7

-2

-2

2

6

25

67

88

-200 -150 -100 -50 0 50 100 150

Wild Rice

Softstem Bulrush

Yellow Waterlily

Subm. Aquatics

River Bulrush

Sand

Pickerelweed

BdLeaved Cattail

Silt

Mixed Emergents

Acres Changed



Figure 5. Number of Buildings in Study Area

SUMMARY

A Geographic Information System was used to replicate 1956 and 1981 analyses ofMerrymeeting Bay wetlands, and extend them 17 years, to 1998. Upland areas for a half-mile buffer surrounding the bay were included. The methods used allow detailed cross-tabulation of conditions between years, showing not only differences in land use andvegetation, but also where those changes occurred.

Trends noted within the Bay in the 1956-1981 interval corroborated the findings of earlierstudies: there was an increase in silt/sand and a decrease in emergent vegetation. Between1981 and 1998 this trend appeared to have stabilized, with no significant increase in silt/sand.Total emergent vegetation remained stable, although there was a shift among species, withwild rice showing an increase. It is difficult to conclude from the data of this study alone,however, whether the dominance of wild rice is a long-term or short-term effect of favorableconditions, since wild rice is an annual.

Number of Buildings in Study Area by Section and Year

0

200

400

600

800

1000

1200

1400

1600

1800

2000

Swan

Island

Eastern

River

Abaga

dasset

R.

Middl

e Bay

Muddy

/Catha

nce Chops

Andro

scoggi

n R.

Section of Study Area

Num

ber

of B

uild

ings

1956

1981

1998



In upland areas, the land use and buildings data clearly reflect the significant increase inhuman population and corresponding decreases in other land use types, notably agriculture.About half of the 1956 agricultural acreage transitioned to forest or a developed condition by1998. The number of buildings in the study area more than tripled between 1956 and 1998.

The digital information resulting from this study will provide a good base for future analyses,as long term trends in the bay continue to be examined.



REFERENCES

Anderson, J. R., Hardy, E. E., Roach, J. T. and Witmer, R. E., 1976, “A Land Use and LandCover Classification System for Use with Remote Sensor Data,” U.S. Geological Survey,Professional Paper 964, p 28, Reston, VA.

Anderson, K. H., 1982. “Summary of Results:1981 Vegetation Survey, Merrymeeting Bay.”Unpub. interdepartmental memorandum, Maine Dept. Inland Fisheries and Wildlife,Augusta, Maine

Cowardin, Lewis M., Carter, Virginia, Golet, Francis C., and LaRoe, Edward T., 1979,“Classification of Wetlands and Deepwater Habitats of the United States,” U.S. Dept. of theInterior, Fish and Wildlife Service, Office of Biological Services, Washington, D.C.

Olson, David P., 1958. “The use of aerial photographs in studies of marsh vegetation.” M.S. Thesis, University of Maine, Orono, Maine

Spencer, Howard E. Jr., 1966. “Merrymeeting Bay Investigations.” Unpub. Manuscript. JobCompletion Report Job A-4, P.R. Proj. W-37-R-14. Maine Dept. Inland Fisheries andWildlife, Augusta, Maine

james w. sewall company april 12, 2000 - merrymeeting bayjames w. sewall company april 12, 2000...

Documents